Driving with Caution About Fully Occluded Areas Based on Occupancy Maps
Yuanxin Zhong, Huei Peng · 2023
Safety is one of the key challenges for autonomous vehicles and correct handling of an occluded environment is a critical safety problem. To get prepared for hidden road users, existing methods usually use high-definition (HD) maps with risk assessment methods to avoid possible crashes. In this paper, we propose a trajectory planning method to avoid crashing into hidden road users based on a grid map that can be generated online. The main contribution is to generate imaginary phantoms to mimic the behavior of real road users. We demonstrated that the proposed method is flexible and effective through experiments. The vehicle with the proposed planning algorithm can navigate through narrow and occluded environments safely and smoothly.